EP1840831A1 - Égalisation d'histogramme adaptative pour des images présentant un fort contraste local - Google Patents

Égalisation d'histogramme adaptative pour des images présentant un fort contraste local Download PDF

Info

Publication number
EP1840831A1
EP1840831A1 EP06006925A EP06006925A EP1840831A1 EP 1840831 A1 EP1840831 A1 EP 1840831A1 EP 06006925 A EP06006925 A EP 06006925A EP 06006925 A EP06006925 A EP 06006925A EP 1840831 A1 EP1840831 A1 EP 1840831A1
Authority
EP
European Patent Office
Prior art keywords
histogram
pixels
histogram equalization
region
adaptive
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP06006925A
Other languages
German (de)
English (en)
Inventor
Zhichun Lei
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sony Deutschland GmbH
Original Assignee
Sony Deutschland GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sony Deutschland GmbH filed Critical Sony Deutschland GmbH
Priority to EP06006925A priority Critical patent/EP1840831A1/fr
Priority to US11/618,030 priority patent/US7822272B2/en
Publication of EP1840831A1 publication Critical patent/EP1840831A1/fr
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/40—Image enhancement or restoration using histogram techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20004—Adaptive image processing
    • G06T2207/20012—Locally adaptive

Definitions

  • the present invention relates to a method, apparatus and computer program product for contrast enhancement of images based on adaptive histogram equalization.
  • it relates to preventing adaptive histogram equalization from causing fading artifacts and object extension artifacts.
  • Histogram equalization is a an effective technique for improving the contrast of an image.
  • the guiding principle of histogram equalization is to equate good contrast of an image with a uniform distribution of the histogram of the intensity of the image's pixels.
  • Histogram equalization proceeds typically as follows. First, a histogram is built from the values of intensity of each pixel of an image. After building, the histogram is advantageously processed by smoothing it. Then a cumulation function is derived from the histogram. This typically is the cumulative histogram. From the cumulation function then, a mapping function is generated. This typically involves normalization of the cumulation function and applying an offset. Finally, the intensity values of the pixels are mapped to new values of intensity as prescribed by the mapping function. The result of this process will be an image with a more uniformly distributed histogram, that is, an image with an enhanced contrast.
  • This process can be applied to the entire image at once or the image can be divided into possibly overlapping subsections called regions or blocks to each of which the process is applied separately.
  • the first approach is called global histogram equalization
  • the later is called local histogram equalization or adaptive histogram equalization.
  • the local histogram equalization can well adapt to local image features, whereas the global histogram equalization can encounter problems in case of such local features.
  • the computational load of local histogram equalization is usually much higher than that of global histogram equalization, it has attracted more and more interest in the area of image contrast enhancement, due to its superior performance.
  • Blocking artifacts that is, the block structure used for block-wise processing is discernible in the processed image, most notably at the block boundaries.
  • Fading artifact which is the term referring to the fact that some areas of an image go pale after adaptive histogram equalization. Typically areas near contrasting objects are affected. Since the human visual perception is much finer for (human) skin colored objects than general objects, this artifact is especially disturbing when skin color areas are affected.
  • Boundary smearing artifact which is the term referring to the fact that boundary areas of contrasting objects (e.g. dark and bright objects) become "smeared" or un-sharp and thicker than before.
  • the boundary smearing artifact is also known as object extension artifact.
  • Overall image contrast loss Since the local processing is unaware of the global histogram information, the global contrast may be negatively affected. For instance, local histogram equalization can make the background of an image as clear as the target object on that image. This emphasized background confuses the human sense of distance.
  • the European patent application EP 1 515 274 which is of the same applicant as is the present invention, addresses the problem of blocking artifacts and overall image contrast loss.
  • Homogenizing histograms is computationally more efficient than overlapping of blocks, a technique also employed to reduce blocking artifacts and the overall image contrast loss. Histogram homogenization can be combined with overlapping blocks.
  • EP 1 465 436 which is of the same applicant as is the present invention, addresses the fading artifact.
  • the method disclosed in EP 1 465 436 builds the histogram from the luminance component of an image signal, such as a TV signal.
  • the chrominance components of the image signal are used to identify skin color areas. Pixels found in skin color areas are used to modify the histogram in a way that it distributes more uniformly, so that the histogram equalization does not change or only slightly changes the signals in human skin color areas. Pixels found in non-skin colored areas are used to build the histogram in the same way as in standard histogram equalization.
  • Object of the invention is to reduce the object extension artifacts and fading artifacts of adaptive histogram equalization.
  • the present invention provides an adaptive histogram equalization method for enhancing the contrast of digital images comprising the steps of dividing an image into regions of pixels, determining structures of local pixel value differences of a predefined strength of the image, building for every region a histogram of the pixel values based on the determined structures of local pixel value differences and mapping pixel values of each region based on the histogram corresponding to the region.
  • an adaptive histogram equalization processor for enhancing the contrast of digital images comprising a region specifying device for dividing an image into regions of pixels, a structure determining device for determining structures of local pixel value differences of a predefined strength of the image, a histogram building device for building for every region a histogram of the pixel values, based on the determined structures of local pixel value differences, a mapping device for mapping pixel values of each region based on the histogram corresponding to the region.
  • the present invention further provides a software product comprising program information that, when executed in a computing device, is operable to carry out the above adaptive histogram equalization method.
  • the apparatus, method and software product for adaptive histogram equalization provided by the present invention reduce fading artifacts and boundary smearing artifacts.
  • the present invention permits the reduction of fading artifacts in skin color areas without requiring access to color information and allows for the reduction of fading artifacts also in non skin color regions.
  • the adaptive histogram equalization method is favorably developed as set forth below:
  • a first side and a second side opposite to said first side in relation to the structure are defined.
  • the histogram is built from the pixel values of the pixels located on said first side of the structure.
  • the amount of pixels located on said second side of the structure modifies the histogram, whereby the larger the number of pixels on said second side the higher the degree of balance or of uniform distribution of the histogram obtained in said building step.
  • the pixels on said second side are accounted for in the histogram as if one part of the pixels were carrying a minimum value and the other part a maximum value.
  • the adaptive histogram equalization method comprises a step of smoothing the histograms or smoothing functions derived from the histograms and used, directly or indirectly, in the step of mapping pixel values.
  • the histogram is built from the pixel values of all pixels of the region.
  • said structures of local pixel value differences of a predefined strength are determined by an edge detection method, for example the Sobel operator.
  • edge detection method determines whether single and sparsely distributed edge points determined by the edge detection method are deleted and/or broken edges are connected by means of e.g. morphological filters.
  • the adaptive histogram equalization method favorably comprises a step of determining skin colored pixels, whereby for regions with skin colored pixels the amount of skin colored pixels modifies the histogram, whereby the larger the number of skin colored pixels in the region the higher the degree of balance or of uniform distribution of the histogram obtained in said building step.
  • the histograms are homogenized.
  • the regions are overlapping and only the values of pixels located in the non-overlapping part of a region are mapped in the step of mapping pixel values of each region based on the histogram corresponding to the region.
  • the pixel values are values relating to the brightness and/or colorness of the pixels, for example the luminance value and/or chrominance or chrominance difference values of a YUV, YIQ, or YCbCr color or other model based image format.
  • the adaptive histogram equalization processor is favorably developed as set forth below:
  • a first side and a second side opposite to said first side in relation to the determined structure are defined.
  • the histogram building device is adapted to build the histogram from the pixel values of the pixels located on said first side of the determined structure.
  • the histogram building device is adapted to modify the histogram based on the amount of pixels located on said second side of determined structure, whereby the larger the number of pixels on said second side the higher the degree of balance or of uniform distribution of the histogram obtained by the histogram building device.
  • the histogram building device is adapted to account for the pixels on said second side in the histogram as if one part of the pixels were carrying a minimum value and the other part a maximum value.
  • the processor further comprises a filtering device for smoothing the histograms or for smoothing functions derived from the histograms and used, directly or indirectly, by the mapping device for the mapping of the pixel values.
  • the histogram building device is adapted to build the histogram from the pixel values of all pixels of the region.
  • the structure determining means is adapted to detect said structures of local pixel value differences of a predefined strength by an edge detection method, for example the Sobel operator.
  • the structure determining means is adapted to delete single and sparsely distributed edge points determined by the edge detection method and/or is adapted to connect broken edges by means of e.g. morphological filters.
  • the processor further comprises a skin color detection device for determining skin colored pixels, whereby for regions with skin colored pixels the histogram building device is adapted to modify the histogram based on the amount of skin colored pixels, whereby the larger the number of skin colored pixels in the region the higher the degree of balance or of uniform distribution of the histogram obtained by the histogram building device.
  • histograms are homogenized.
  • the regions are overlapping and the mapping device is adapted to map only the values of pixels located in the non-overlapping part of a region.
  • the pixel values are values relating to the brightness and/or colorness of the pixels, for example the luminance value and/or chrominance or chrominance difference signals of a YUV, YIQ, or YCbCr or other color model based image format.
  • the present invention is most easily implemented if the input image is digital. If this is not the case, conversion from analog format to digital format can be affected prior to the contrast enhancement processing of the image.
  • the input image may be a color or a monochrome image.
  • the input image may be a single image or it may be an image sequence. In other words, the present invention may be applied to still images like photographs or to image sequences like video images.
  • the present invention can for example be employed when the input image format comprises data which relates to the brightness of the image's pixels, for example the grayscale value of a grayscale scanned image or the luminance of a video signal, e.g. of a standard TV-signal.
  • Such brightness data may however be generated from the data of the input format any time before the brightness data is used, e.g. by a conversion from Red-Green-Blue (RGB) color model to a Luminance-Chrominance color model, like the YUV, YIQ and YCbCr color models used by the PAL, NTSC and video "TV-standards", respectively.
  • RGB Red-Green-Blue
  • Luminance-Chrominance color model like the YUV, YIQ and YCbCr color models used by the PAL, NTSC and video "TV-standards", respectively.
  • the green component of a color signal has the advantage that the green color component normally has the best signal to noise ratio.
  • the present invention may be applied to enhance the contrast of data which is not data relating to brightness, this may be data relating to color or any other data.
  • the invention can likewise be seen in method or in a corresponding apparatus or computer program product. While the present description may, for the sake of brevity, limit itself to a description of the method of the invention, the teachings of this specification are to be understood as applying equally to and encompassing a corresponding apparatus and computer program product capable of effecting the functionality of the described method.
  • Fig. 1 shows a block diagram of an adaptive histogram equalization method according to the present invention.
  • a step S2 an image is divided into regions of pixels.
  • a step S4 structures of local pixel value differences of a predefined strength of the image are determined.
  • a step S8 for every region, a histogram of the pixel values is built, based on the determined structures of local pixel value differences.
  • pixel values of each region are mapped based on the histogram corresponding to the region.
  • skin colored pixels are determined in step S6. In that case, the building of the histogram in step S4 is also based on the determined skin colored pixels.
  • the histogram of each region is smoothed in a step S10.
  • Fig. 2 shows a block diagram of a corresponding adaptive histogram equalization processor 1.
  • the adaptive histogram equalization processor 1 comprises a region specifying device 2 for dividing an image into regions, a structure determining device 4 for determining structures of local pixel value differences of a predefined strength of the image, a histogram building device 8 for building, for every region, a histogram of the pixel values, based on the determined structures of local pixel value differences and a mapping device 12 for mapping pixel values of each region based on the histogram corresponding to the region.
  • the processor 1 comprises a skin color detection device 6 for determining skin colored pixels.
  • the histogram building device 8 is adapted to build the histogram based on the determined structures of local pixel value differences and based on the determined skin colored pixels.
  • the processor 1 comprises a filtering device 10 for smoothing the histograms.
  • the devices 2, 4, 6, 8, 10, 12 of the processor 1 can be implemented by any computing device or any device capable of information processing. For example, by a suitable programmed personal computer, by a computing device based on a digital signal processor adapted for image manipulation, or any other microprocessor based information processing device.
  • the devices 1, 4, 6, 8, 10, 12 of the processor 1 may be implemented by the very same piece of hardware or by different pieces of hardware. An embodiment of an adaptive histogram equalization method according to the present invention is now explained with reference to Fig. 3 to 7.
  • processor 1 may require additional elements including, for example, housing, power supply, data interfaces, data connections, buffer. Such elements needed for practical implementation are not mentioned in the description since they are a matter of course for the skilled person set out to implement the present invention.
  • the input image processed by the embodiment of the method is assumed to be an image in a digital format, the format comprising luminance data.
  • the regions of pixels are of rectangular shape and called block, see below for more details on the division of the image into regions of pixels.
  • other than rectangular forms can be employed too.
  • the block-wise processing is from left to right and from top to bottom. This corresponds to the order in time in which standard TV and video-signals carry the image information.
  • To process a signal in the inherent order of the signal is favorable since processing is faster and less memory is required.
  • Many other ways of block-processing are possible though, for example, blocks could be, partially at least, processed in parallel.
  • FIG. 3 a flow diagram of a central part of the method's embodiment is shown.
  • the diagram comprises the processing of a block from the determination of structures of local pixel value differences up to the processing of the histogram including smoothing of the histogram.
  • step S20 structures of local luminance value difference of a predefined strength of a block are determined by an edge detection method.
  • the sensitivity of the edge detection is adjustable via an edge threshold parameter.
  • the predefined strength defines a lower limit for the strength of the local pixel value differences whereby strength can be measured by any appropriate method.
  • typically contrasting objects are the cause of fading artifacts and object extension artifacts.
  • the determined local pixel value differences can be seen as large or strong compared to the local pixel value differences corresponding to a gradual change. In this sense, the structures determined are structures of strong local pixel value differences. Though, nothing prevents the present applications from being used with low pixel value differences.
  • a more precise value for the strength of the local pixel value differences can hardly be given.
  • Image properties such as the amount of noise in the image or the kind of motif of the image have an influence on the finding of an operable value.
  • this depends on the personal taste of an operator, in high quality video post-production, for example, the edge parameter may be adjusted according to needs by a human operator.
  • the user may have direct access to the strength.
  • the user of a camcorder may select from a range of presets each optimized for typical recording situations, whereby the preset, inter alia, sets a value for the strength of the pixel value differences to be determined.
  • Another possibility is an automatic setting based on recording conditions such as for example lightning condition or based on other previously gained information about the image, for example, information gained in a previous image manipulation step.
  • different values might be set for different areas of the image, for example, for different regions.
  • the information obtained can be displayed as an image.
  • the image of Fig. 6 was obtained processing the entire image of Fig. 5 using the well-known Sobel operator.
  • the determined structures of local pixel value difference are displayed white on a black background.
  • the determined structures of local pixel value difference may also be called edges, edge lines, edge points, boundary, boundary line, boundary information, etc.
  • edges, edge lines, edge points, boundary, boundary line, boundary information, etc. One sees that the structures of local pixel value difference match with the boundary lines of contrasting objects.
  • the structures of local pixel value differences are not limited to any shape.
  • step S22 when coming in from step S20, it is determined if edges exist in the block. If no, the histogram proceeds to step S30. If, yes the histogram proceeds to step S24.
  • step S24 the pixels behind the edge are accounted for in the histogram as if alternatively having a minimum and a maximum luminance value, e.g. 16 and 235 for standard 8 bits quantized TV-signal.
  • the minimum and the maximum value may differ to some extent from the minimally and maximally possible values in a given image format as long as they still correspond to a rather small or rather large value within the possible dynamic range.
  • step S26 it is decided which kind of block end is reached. If the block end in horizontal x-direction is reached the process proceeds to step S28 where the process locates to the next line and from where the process proceeds to step S22. If the block end in vertical y-direction is reached, the process proceeds to step S30.
  • step S22 when coming in from step S 28, it is decided whether further edges exist. If yes, the process proceeds to step S24. If no, the process proceeds to step S30.
  • step S30 the pixel values of the pixels lying in front of the edges are directly taken into account for building the histogram. After step S30 the histogram building is finished and the process proceeds to step S32.
  • the histogram is built as usual. For areas with strong local contrast, only the pixels in front of the detected object boundary are directly taken into account for the histogram building. The occurrences of other pixels are not directly taken into account but are used as indices to modify the histogram in question to make it more uniform, or have it behave like a more uniform histogram, here by alternatively counting a minimal and a maximal value to the histogram until the block end is reached. The alternatively counting is done until both the horizontal (x) and the vertical (y) block end is reached. From the edge starting position to the end of the block, the histogram takes only the minimal and maximal value.
  • step S32 the histogram is processed. This includes smoothing the histogram, for example by filtering it with a low pass filter.
  • the histogram obtained may be mixed in a weighted sum with the previously obtained histograms of the adjacent blocks.
  • This is a simple scheme for histogram homogenization.
  • histogram homogenization reduces artifacts from being caused by single edge points, sparsely distributed edge points or broken edges.
  • a mapping function is derived from a cumulative function of the smoothed histogram and the luminance values of the block are mapped according to the mapping function. Note however, that the present embodiment does not map all pixels of a block. This is explained in more detail below. When all blocks have been processed and the corresponding pixels have been mapped the histogram equalization is finished for the input image.
  • the pixels in front of a boundary line are taken directly into account, for building the histogram, whereas the pixels after a boundary line are used as indicators to modify the histogram in question to make it more uniform, or have it behave like a more uniform histogram.
  • this is effected by alternately counting a minimum and a maximum luminance value until the end of the block is reached.
  • This balances the histogram which results in a histogram more uniformly distributed after smoothing than a corresponding smoothed histogram built entirely from the actual pixel values would be.
  • the effect of balancing is a histogram that behaves like a more uniformly distributed histogram.
  • Other ways are possible to modify the building of a histogram to achieve a similar result however.
  • uniformity can be achieved more directly.
  • the pixels which are not accounted for with their actual values could for example be used to fill the gaps in a histogram built from the other pixels, which are accounted for with their actual values, thereby adding directly to the uniformity of the histogram.
  • the histogram modification does not play or play only a little role in the final result.
  • fading artifacts and object extension artifacts are usually very weak, that is, are invisible anyway.
  • the less the part that the pixels in front of the boundary line take up of the block the more apparently fading artifacts and object extension artifacts appear according to experiences, however, the stronger is also the histogram modification and with it the suppression of artifacts.
  • the amount of artifact reduction is dosed according to the actual need.
  • the pixels in front of a boundary line are used as indicators to modify the histogram in question to make it more uniform or have it behave like a more uniform histogram and the pixels behind the boundary line are directly taken into account for building the histogram.
  • Block overlapping is known to reduce blocking artifacts. Block overlapping, however, may also aggravate fading and object extension artifacts.
  • Fig. 4 shows three rectangular blocks overlapping densely in a horizontal x-direction. For better visualization, these three blocks are also shown separately (the three lower blocks). In order not to map twice, only the pixels in the area filled with a pattern should be mapped with the mapping function derived from the entire block.
  • the block structure used by the embodiment overlaps also in vertical direction. For simplicity, the vertical overlapping of blocks is not shown in Fig. 4.
  • the area that is mapped with the mapping function must be further reduced. Only the non-overlapping area, in horizontal as in vertical direction, is mapped with the mapping function derived from the entire block. Thus only a very small part of the pixels of each block is mapped, although the histograms are built from the entire blocks. Since only the top left pixels are mapped, the histogram modification usually does not affect the pixels behind the detected boundary line.
  • Fig. 7 two difference pictures between processed and original image are displayed (added by an offset and amplified by a factor for better visualization using printed pictures).
  • the original image was processed using the embodiment of the histogram equalization method according to the present invention.
  • the original image was processed using prior art histogram equalization.
  • the left (prior art) difference picture brightness variations corresponding to pronounced object extension artifacts are present at the boundary line discerning shoulders and arms from the background: In the boundary area the pullover got darker and the background got lighter by processing the image with the prior art technique.
  • brightness variations corresponding to fading artifacts are present at the boundary line between pullover and neck: The skin color got lighter close to the pullover.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Facsimile Image Signal Circuits (AREA)
  • Image Processing (AREA)
EP06006925A 2006-03-31 2006-03-31 Égalisation d'histogramme adaptative pour des images présentant un fort contraste local Withdrawn EP1840831A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP06006925A EP1840831A1 (fr) 2006-03-31 2006-03-31 Égalisation d'histogramme adaptative pour des images présentant un fort contraste local
US11/618,030 US7822272B2 (en) 2006-03-31 2006-12-29 Adaptive histogram equalization for images with strong local contrast

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
EP06006925A EP1840831A1 (fr) 2006-03-31 2006-03-31 Égalisation d'histogramme adaptative pour des images présentant un fort contraste local

Publications (1)

Publication Number Publication Date
EP1840831A1 true EP1840831A1 (fr) 2007-10-03

Family

ID=36569142

Family Applications (1)

Application Number Title Priority Date Filing Date
EP06006925A Withdrawn EP1840831A1 (fr) 2006-03-31 2006-03-31 Égalisation d'histogramme adaptative pour des images présentant un fort contraste local

Country Status (2)

Country Link
US (1) US7822272B2 (fr)
EP (1) EP1840831A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2053558A2 (fr) 2007-10-26 2009-04-29 Qualcomm Incorporated Amélioration de la qualité d'image avec diffusion d'histogramme
CN102456222A (zh) * 2010-10-29 2012-05-16 深圳迈瑞生物医疗电子股份有限公司 图像中组织均衡的方法和装置
CN101408973B (zh) * 2007-10-10 2012-09-26 奇景光电股份有限公司 图像处理方法及其装置
CN117788468A (zh) * 2024-02-26 2024-03-29 江西福松和安医疗科技有限公司 喉部图像处理方法、系统、喉镜、可调节气道建立装置

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8154663B2 (en) * 2007-01-16 2012-04-10 Sigma Designs, Inc. System and method for adaptive contrast enhancement of video signals
WO2009078957A1 (fr) 2007-12-14 2009-06-25 Flashfoto, Inc. Systèmes et procédés de segmentation basée sur des règles d'objets en vue de face complète ou partielle sur des images en couleurs
US8442267B2 (en) * 2008-12-24 2013-05-14 Electronics And Telecommunications Research Institute Apparatus and method for detecting upper body posture and hand posture
US8411986B2 (en) * 2009-04-13 2013-04-02 Flashfoto, Inc. Systems and methods for segmenation by removal of monochromatic background with limitied intensity variations
US9524700B2 (en) 2009-05-14 2016-12-20 Pure Depth Limited Method and system for displaying images of various formats on a single display
US8928682B2 (en) * 2009-07-07 2015-01-06 Pure Depth Limited Method and system of processing images for improved display
US8670615B2 (en) * 2009-09-30 2014-03-11 Flashfoto, Inc. Refinement of segmentation markup
CN101739672B (zh) * 2009-12-02 2015-08-05 北京中星微电子有限公司 一种基于分区域插值的直方图均衡方法和装置
US8606009B2 (en) 2010-02-04 2013-12-10 Microsoft Corporation High dynamic range image generation and rendering
US9311567B2 (en) 2010-05-10 2016-04-12 Kuang-chih Lee Manifold learning and matting
JP5147903B2 (ja) * 2010-07-12 2013-02-20 キヤノン株式会社 画像処理装置、画像処理方法、及びプログラム
JP5991486B2 (ja) * 2010-08-04 2016-09-14 日本電気株式会社 画像処理方法、画像処理装置及び画像処理プログラム
WO2013003335A1 (fr) * 2011-06-27 2013-01-03 Vytran, Llc Appareil et procédés de détermination d'un axe de biréfringence d'une fibre optique à maintien de polarisation
US9311740B2 (en) 2012-03-27 2016-04-12 Carestream Health, Inc. Method for enhancing reconstructed 3-D tomosynthesis volume image
KR102109933B1 (ko) 2013-07-01 2020-05-12 삼성전자주식회사 디스플레이의 곡률을 참조하는 영상 처리장치 및 방법
KR102146560B1 (ko) 2014-02-17 2020-08-20 삼성전자주식회사 영상 보정 방법 및 장치
MA49853A (fr) * 2017-08-07 2021-04-21 Imago Systems Inc Système et procédé de visualisation et de caractérisation d'objets dans des images
US10977811B2 (en) * 2017-12-20 2021-04-13 AI Analysis, Inc. Methods and systems that normalize images, generate quantitative enhancement maps, and generate synthetically enhanced images
KR102368229B1 (ko) 2018-02-06 2022-03-03 한화테크윈 주식회사 영상처리장치 및 방법
US11200653B2 (en) 2019-08-06 2021-12-14 Samsung Electronics Co., Ltd. Local histogram matching with global regularization and motion exclusion for multi-exposure image fusion
WO2021025445A1 (fr) * 2019-08-06 2021-02-11 Samsung Electronics Co., Ltd. Mise en correspondance d'histogramme local avec régularisation globale et exclusion de mouvement pour fusion d'image à expositions multiples
CN114898117A (zh) * 2022-05-25 2022-08-12 南京邮电大学 一种基于均衡化结构纹理表征的序列红外图像目标检测方法

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1515274A1 (fr) * 2003-07-07 2005-03-16 Sony International (Europe) GmbH Procédé et dispositif pour obtenir des histogrammes d'image

Family Cites Families (23)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3983320A (en) * 1975-08-25 1976-09-28 Hughes Aircraft Company Raster display histogram equalization
US4187519A (en) * 1978-08-17 1980-02-05 Rockwell International Corporation System for expanding the video contrast of an image
US5042077A (en) * 1987-10-02 1991-08-20 General Electric Company Method of highlighting subtle contrast in graphical images
US5524070A (en) * 1992-10-07 1996-06-04 The Research Foundation Of State University Of New York Local adaptive contrast enhancement
US5426684A (en) * 1993-11-15 1995-06-20 Eastman Kodak Company Technique for finding the histogram region of interest for improved tone scale reproduction of digital radiographic images
US5588071A (en) * 1994-10-26 1996-12-24 Minnesota Mining And Manufacturing Company Identifying an area of interest using histogram data arranged in predetermined sequence
EP0712092A1 (fr) * 1994-11-10 1996-05-15 Agfa-Gevaert N.V. Procédé d'amélioration d'images
WO1996016534A2 (fr) * 1994-11-25 1996-06-06 Sophisview Technologies, Ltd. Systeme et procede de diagnostic des affections des tissus vivants
US5982926A (en) * 1995-01-17 1999-11-09 At & T Ipm Corp. Real-time image enhancement techniques
KR980003998A (ko) * 1996-06-27 1998-03-30 김광호 제한된 분포의 히스토그램 변환을 이용한 화질 개선방법
US6049626A (en) * 1996-10-09 2000-04-11 Samsung Electronics Co., Ltd. Image enhancing method and circuit using mean separate/quantized mean separate histogram equalization and color compensation
FI104521B (fi) * 1997-01-30 2000-02-15 Nokia Multimedia Network Termi Menetelmä kuvasekvenssien kontrastin parantamiseksi
KR100261214B1 (ko) * 1997-02-27 2000-07-01 윤종용 영상처리 시스템의 콘트라스트 확장장치에서 히스토그램 등화방법 및 장치
US6282307B1 (en) * 1998-02-23 2001-08-28 Arch Development Corporation Method and system for the automated delineation of lung regions and costophrenic angles in chest radiographs
US6650774B1 (en) * 1999-10-01 2003-11-18 Microsoft Corporation Locally adapted histogram equalization
US6804408B1 (en) * 1999-12-22 2004-10-12 Eastman Kodak Company Method for enhancing a digital image with noise-dependent control of texture
US6915024B1 (en) * 2000-09-29 2005-07-05 Hewlett-Packard Development Company, L.P. Image sharpening by variable contrast mapping
US6836570B2 (en) * 2001-11-14 2004-12-28 Eastman Kodak Company Method for contrast-enhancement of digital portal images
EP1345172A1 (fr) * 2002-02-26 2003-09-17 Sony International (Europe) GmbH Amélioration de contraste d'images numériques
EP1465436A1 (fr) 2003-04-03 2004-10-06 Sony International (Europe) GmbH Préservation d'impression naturelle pour des images à histogramme egalisé
US7221408B2 (en) * 2003-08-15 2007-05-22 Samsung Electronics Co., Ltd. Adaptive contrast enhancement method for video signals based on time-varying nonlinear transforms
US7573533B2 (en) * 2004-10-15 2009-08-11 Genesis Microchip Inc. Method of generating transfer curves for adaptive contrast enhancement
US7636496B2 (en) * 2006-05-17 2009-12-22 Xerox Corporation Histogram adjustment for high dynamic range image mapping

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1515274A1 (fr) * 2003-07-07 2005-03-16 Sony International (Europe) GmbH Procédé et dispositif pour obtenir des histogrammes d'image

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JIA-GUU LEU: "IMAGE CONTRAST ENHANCEMENT BASED ON THE INTENSITIES OF EDGE PIXELS", CVGIP GRAPHICAL MODELS AND IMAGE PROCESSING, ACADEMIC PRESS, DULUTH, MA, US, vol. 54, no. 6, 1 November 1992 (1992-11-01), pages 497 - 506, XP000332318, ISSN: 1077-3169 *
PARANJAPE R B ET AL: "ADAPTIVE-NEIGHBORHOOD HISTOGRAM EQUALIZATION FOR IMAGE ENHANCEMENT", CVGIP GRAPHICAL MODELS AND IMAGE PROCESSING, ACADEMIC PRESS, DULUTH, MA, US, vol. 54, no. 3, 1 May 1992 (1992-05-01), pages 259 - 267, XP000281965, ISSN: 1077-3169 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101408973B (zh) * 2007-10-10 2012-09-26 奇景光电股份有限公司 图像处理方法及其装置
EP2053558A2 (fr) 2007-10-26 2009-04-29 Qualcomm Incorporated Amélioration de la qualité d'image avec diffusion d'histogramme
EP2053558A3 (fr) * 2007-10-26 2009-09-09 Qualcomm Incorporated Amélioration de la qualité d'image avec diffusion d'histogramme
WO2009055727A3 (fr) * 2007-10-26 2009-10-15 Qualcomm Incorporated Amélioration de la qualité d'images par diffusion d'histogrammes
CN102456222A (zh) * 2010-10-29 2012-05-16 深圳迈瑞生物医疗电子股份有限公司 图像中组织均衡的方法和装置
CN102456222B (zh) * 2010-10-29 2013-12-04 深圳迈瑞生物医疗电子股份有限公司 图像中组织均衡的方法和装置
CN117788468A (zh) * 2024-02-26 2024-03-29 江西福松和安医疗科技有限公司 喉部图像处理方法、系统、喉镜、可调节气道建立装置
CN117788468B (zh) * 2024-02-26 2024-04-30 江西福松和安医疗科技有限公司 喉部图像处理方法、系统、喉镜、可调节气道建立装置

Also Published As

Publication number Publication date
US7822272B2 (en) 2010-10-26
US20070230788A1 (en) 2007-10-04

Similar Documents

Publication Publication Date Title
US7822272B2 (en) Adaptive histogram equalization for images with strong local contrast
EP0723364B1 (fr) Procédé d'amélioration d'image et temps réel
KR101625362B1 (ko) 입력 이미지 데이터를 출력 이미지 데이터로 변환하기 위한 방법, 입력 이미지 데이터를 출력 이미지 데이터로 변환하기 위한 이미지 변환 유닛, 이미지 프로세싱 장치, 디스플레이 디바이스
US10565742B1 (en) Image processing method and apparatus
US6094508A (en) Perceptual thresholding for gradient-based local edge detection
CN111915534B (zh) 一种图像去噪的方法及装置
US20110188776A1 (en) Image processing method and image processing apparatus
US7612784B2 (en) Image processor and method, computer program, and recording medium
JP2007310886A (ja) 画像データの自動マッピング方法及び画像処理デバイス
JP2006121713A (ja) コントラストの強調
US20100188531A1 (en) Image detail enhancement
US20130064445A1 (en) Adaptive Dithering During Image Processing
Smith et al. Beyond tone mapping: Enhanced depiction of tone mapped HDR images
CN111739041B (zh) 图像边框的裁剪方法、装置及设备
CN114841904A (zh) 图像融合方法、电子设备及存储装置
JP2009010636A (ja) 適応ヒストグラム等化方法及び適応ヒストグラム等化装置
EP1183408B1 (fr) Amelioration d'un signal d'image
Sun et al. Modified luminance based adaptive MSR
CN111915497A (zh) 一种图像黑白增强方法、装置、电子设备和可读存储介质
CN105654541B (zh) 窗口图像处理方法和装置
JPH10340332A (ja) 画像処理装置、画像処理方法、画像処理制御プログラムを記録した媒体
JP2003085553A (ja) ノイズリダクション回路
JP3501151B2 (ja) 画像処理装置、画像処理方法、画像処理制御プログラムを記録した媒体
WO2024060360A1 (fr) Procédé et appareil de fusion d'images, dispositif électronique et support de stockage
JP3450144B2 (ja) 画像ノイズ分析方法

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC NL PL PT RO SE SI SK TR

AX Request for extension of the european patent

Extension state: AL BA HR MK YU

17P Request for examination filed

Effective date: 20080103

17Q First examination report despatched

Effective date: 20080206

AKX Designation fees paid

Designated state(s): DE FR GB

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20080617